Keywords:
ego networks

References:
M. Everett and S. P. Borgatti. Ego network betweenness. Social Networks, 27(1):31--38, 2005.

Full Text:
... (v. 2005-07-01). Downloaded fromhttp://crawdad.cs.dartmouth.edu/mit/reality.[4] M. Everett and S. P. Borgatti. Ego network betweenness.Social Networks, 27(1):31?38, 2005.[5] S. Farrell and V. Cahill. Delay- ...

Keywords:
ego networks

Abstract:
... propose an unsupervised approach to automatically detect circles in an ego network such that each circle represents a densely knit community of ...

References:
J. J. McAuley and J. Leskovec. Learning to discover social circles in ego networks. In NIPS, pages 548--556, 2012.
J. J. McAuley and J. Leskovec. Discovering social circles in ego networks. TKDD, 8(1):4, 2014.

Full Text:
... propose an unsupervised approach to au-tomatically detect circles in an ego network such that eachcircle represents a densely knit community of researchers.Our ... DescriptorsH.2.8 [Database Application]: Data mining; E.1 [DataStructure]: Graphs and networksKeywordsCircles; ego networks; ; prediction?First two authors have equal contributions.Permission to make digital ... anypoint of time.Figure 1: (Color online) A hypothetical exampleshowing an ego network of an author u with labeledcircles. Alters might belong to ...
... linked circles. In practice, sincethe topological evidence in such small ego networks is less,the traditional community finding algorithms fail to discovermeaningful circles ...
et al. [17] attempted topredict the h-index evolution through ego networks, , observ-ing that this factor increases if one can choose ... Leskovec [15, 16] were the first who exploredsocial circles in ego networks. . They mapped this problemas a multi-membership node clustering problem ... because of the dearth of informationon topological structure in the ego network of each author[9, 16]. According to them, under such circumstances ... EGO-CENTRIC CIRCLESOur model for detecting ego-centric circles applies to anygeneral ego network, , where each node is considered as anego and the ... set of alters. The ego is said to spawn the ego network, ,but is not considered as a part of the network. ... of the network. Our methodintends to discover circles in this ego network in an unsuper-vised fashion, leveraging properties specific to nodes as ... referred to as the center node, is responsible forspawning the ego network, , but does not itself feature as apart of the ... not itself feature as apart of the network. So the ego network of u is essentiallythe subgraph induced by the alters of ...
... to each node (author) presentin our dataset, thus obtaining 821,633 ego networks. . Anillustrative example of an ego network is shown in Figure 1.However, in this experiment we consider ... ego and exclude the ego and itsattached edges from the ego network, , as mentioned earlier.6. FEATURE EXTRACTIONProfile information of each author ... publications.The general features capture independent characteristicsof each author in the ego network and are listed below:? The normalized number of citations the ...
... circle-specificanalysis.8.1 Author-specific AnalysisHere we study how the circles in the ego networks of thehighly-cited authors differ from those of the low-cited au-thors. ... Figure 4(a) shows the number of ego-centric circlesappearing in the ego network of each author. Note thatin all of the experiments, we ... have high number of collaborators,the number of alters in their ego networks is also high, andthus more number of ego-centric circles are ... that since highly-citedauthors have more number of alters in their ego networks, ,each of them also belongs to multiple local circles of ... we measure the average cliquishness of hersurrounding circles in the ego network. . Figure 4(d) showsthat the average value of cliquishness initially ...
... each group also representsthe time of collaboration. For instance, the ego network ofAuthor 12 (one of the highly-cited authors) is shown in ...
ego network encircles authorsin Data Mining who have coauthored with her during ... log(FC,i) (16)Figure 6(b) captures the average homogeneity of circles inthe ego- -network of authors in four different citation ranges.We note that the ...
... Each feature is appropriately normalized116Figure 8: (Color online) Examples of ego- -networks from three citation zones. Individual nodes have differentcolors corresponding to ...
Leskovec. Learning to discoversocial circles in ego networks. . In NIPS, pages 548?556,2012.[16] J. J. McAuley and J. ... 548?556,2012.[16] J. J. McAuley and J. Leskovec. Discovering socialcircles in ego networks. . TKDD, 8(1):4, 2014.[17] C. McCarty, J. Jawitz, A. Hopkins, ...

Title:
Ego-net digger: a new way to study ego networks in online social networks

Keywords:
ego networks

Abstract:
... <i>ego-net digger</i>, a novel Facebook application for the analysis of ego networks in OSN. Ego-net digger collects users' social data and gives ... acquisition campaign, finding interesting similarities between the structure of Facebook ego networks and the properties found in the anthropology literature. Specifically, we ... is a clear evidence of the presence of the same ego network structure - i.e., the Dunbar's circles - as found in ...

Full Text:
images/freq_all.epsEgo-net Digger: a New Way to Study Ego Networks inOnline Social NetworksMassimiliano La GalaIIT-CNRvia G. Moruzzi,1Pisa (PI), Italym.lagala@iit.cnr.itValerio ArnaboldiIIT-CNRvia ... there is aclear evidence of the presence of the same ego network struc-ture - i.e., the Dunbar?s circles - as found in ... 12?16, 2012, Beijing, China.Copyright 2012 ACM 978-1-4503-1549-4 ...$10.00.KeywordsOnline Social Networks, Ego networks, , Tie Strength1. INTRODUCTIONThe worldwide proliferation of online social networks(hereinafter ... studied usinga model for the representation of personal social networks,called ego network - i.e., a social network formed of an in-dividual (ego) ... people with whom ego is in contact(alters).While the properties of ego networks in the real worldhave been deeply studied in the anthropology ... been deeply studied in the anthropology and sociolog-ical literature, OSN ego networks are not yet completelyunderstood. Specifically, there is a lack of ... In addition, the fundamental differ-ences between the properties of OSN ego networks and thewell known results about social networks formed in the ... networks formed in the realworld (hereafter referred to as human ego networks) ) are stillunder investigation.In this paper we focus on the ... they show the same structure found in humanego networks. Human ego networks presents a hierarchicalstructure formed of a series of four concentric ... as this frequency decrease. Further detailson the structure of human ego networks are given in Sec-tion 2.1.The study of these social aspects ...
... the users a graphical repre-sentation of the structure of their ego- -networks, , arrangingtheir friends according to the Dunbar?s circles model. Thesecond ... Facebook using public API. To havea complete view of users? ego networks, , we also collect ex-plicit evaluations of their social relationships ... weare able to perform accurate analysis on the properties ofFacebook ego networks, , and to depict their structure as agraphical output for ... the Dunbar?s circles principle)is similar to that found in human ego networks. . To do so weanalyse a small sample of social ... an interesting indica-tion of the similarity of the structure of ego networks in OSNand in real life. Specifically, we find that the ... of the existing work in literature re-garding the properties of ego networks in real life and someof the preliminary results found in ... concept and to show the similarities we have found be-tween ego networks structure in Facebook and in real life.Hence, in Section 5 ... give an overview of the background workin literature on social ego networks analysis from two dif-ferent standpoints. Firstly, we introduce the relevant ... the two concepts appear tobe tightly correlated [3, 7].Studying human ego networks, , anthropologists found thatan individual can maintain only a limited ...
... therefore thenumber of active relationships is also limited.The portion of ego network containing active relationshipsis called active network. In human ego networks the averageactive network size is around 150, the so called ... hypothesis has been validatedby different studies on humans [12, 7].Human ego networks show a typical hierarchical structureformed of four concentric subgroups of ... [1, 5], as a firstproof of the similarity of the ego networks inside physicaland cyber environments. Although these interesting find-ings, the relation ... interesting find-ings, the relation between tie strength and the propertiesof ego networks is not yet completely understood. The de-sign of measures of ... (ii) study the structure and other important prop-erties of OSN ego networks and the relation between humanbehaviour in the virtual and the ... data between theuser and her friends, giving as output the ego network of theuser partitioned into its Dunbar?s circles. Ego-net digger isthe ...
... goals:? give a graphical representation of the Dunbar?s circlesof users? ego networks through the analysis of Facebookdata. This represents an immediate feedback ...
... a web-based interface thatdisplays to the user her Facebook active ego network dividedinto four different groups, according to the Dunbar?s circlesconcept (see ... the tiestrength evaluations in real and virtual worlds.4.1 Analysis of Ego Network StructureThe structure of Dunbar?s circles in human ego networksis based ...
... number of likes sent by ego to alters,calculated for each ego network, , then averaged for all thenetworks. This result indicates the ... contact we should be able tocharacterise the structure of our ego networks, , maintainingan analogy with the results in the anthropology literature.This ... to whom ego sendsdirect replies to tweets.We first consider the ego networks identified by using thefrequency of outgoing communications as the strength ...
... CCDF is similar tothat presented in figure 3To check whether ego networks can be divided into dif-ferent circles - according to the ... nu-merical results, although the analysis in [1] also suggests ahierarchical ego network structure.ent data spaces. The former is compesed of one dimensionalpoints ... representing the values of the frequency of outgoingcommunication of each ego network, , while the latter con-tains the points representing the AdjFreq ... the points representing the AdjFreq index. We verifythat the Facebook ego networks of our sample show a typi-cal number of circles similar ... the k-meansalgorithm with an increasing value of k on each ego network, ,until the variance explained by the partitioning scheme re-mains under ... median 4. Hence, we re-apply the k?means algorithmon all the ego networks fixing k to be equal to 4, using boththe frequency ... (14.03) (41.9) (132.5)dex) is lower than that found in human ego networks. . Thiscould be ascribed to the fact that online ego networks repre-sent only a partial subset of the ego networks of a person inreal life and on-line ego networks are currently still in the in-fancy and in a growing ... in-fancy and in a growing phase (the size of an ego network alsodepends on how long the ego has been active in ... been active in OSN). De-spite this, the structure of Facebook ego networks presentsthe same hierarchical pattern of the structure of human egonetworks. ... results confirm the structure similaritybetween the Dunbar?s circles of the ego networks obtainedby ego-net digger and those found in human real ego ...
the frequency of outgoing and incoming communication,produces an ego network structure closer to that found inthe anthropology literature.The typical frequency ... twice a year (active network). Thestructure we find in Facebook ego networks is thus compat-ible with that found in human ego networks. .4.2 Comparison Between Virtual and RealEgo NetworksIn addition to the ... in the two worlds and we study how muchthe active ego networks in OSN and in real life overlap. Tothis aim we ... present ego-net digger, a novel Facebookapplication for the study of ego networks in Facebook. Ego-net digger, in addition to be a valuable ... output a graphical repre-sentation of the Dunbar?s circles of users? ego networks. . Weanalyse a sample data set obtained using ego-net digger,showing ... prop-erties of OSN. We will also continue studying the relationbetween ego networks in OSN and in real life, analysing howOSN variables could ...

Keywords:
Ego networks

Abstract:
... based on the focal individual's network structure, also called an ego network. . The novel factor in this approach is that it ...

References:
McAuley, J. and Leskovec, J. "Learning to Discover Social Circles in Ego Networks," NIPS, pp. 548--556, 2012

Full Text:
... or friendand then the codecide which inturn, based on than ego network. . Tmake recommendorder would maknext recommendn our experimenusing our confidorder, ...
... al. [6], in which there are five files from every ego network. . T We construct ego?s friends and features using edge ... choice of confidence score from the friendship circles: Since each ego network has more than two friendship circles, the problem is how ... circles. 2. SVM-score order vs. fixed order: We use an ego networks to compare the accuracy of the SVM-score order and fixed ...
... obtained using data from a social network. We use the ego network structure with the ego classifications of circles based on his/her ...
... J. and Leskovec, J. ?Learning to Discover Social Circles in Ego Networks, ,? NIPS, pp.548-556, 2012 [6] Inuzuka, N., Takeuchi, S. and ...

Title:
User profiling in an ego network: co-profiling attributes and relationships

Keywords:
ego network

References:
J. McAuley and J. Leskovec. Learning to discover social circles in ego networks. In P. Bartlett, F. Pereira, C. Burges, L. Bottou, and K. Weinberger, editors, Advances in Neural Information Processing Systems 25, pages 548--556. 2012.

Full Text:
fp862-rui.dviUser Profiling in an Ego Network: : Co-profiling Attributesand Relationships ?Rui Li?, Chi Wang?, Kevin Chen-Chuan ... Management]: Database Applications - DataMining; H.4.0 [Information Systems]: GeneralKeywordsSocial Network; Ego Network; ; User Pro?ling1. INTRODUCTIONUser pro?ling, which refers to the process ...
... co-pro?ling.Speci?cally, in this paper, we focus on co-pro?ling uponevery user?s ego network, , which means that we pro?le auser?s attributes and the ... attributes and the relationship types of his connec-tions given his ego network. . An ego network is a networksurrounding an individual user, called the ego, which ... his friends and thoseamong his friends. We focus on the ego network of each userinstead of a ?full? social network due to ... social network is unavailable to public ingeneral, whereas a user?s ego network ... is often accessible viacertain protocols (e.g., user-authorized applications). Sec-ond, an ego network keeps important connections related toone user and typically contains hundreds ... to evaluate our algorithm on a large col-lection of real-world ego networks. . Speci?cally, we evaluateour proposed algorithm for two tasks, i) ...
... two key di?erences from the above studies.First, we focus on ego- -network, , while the above methodsfocus on either a ?full? social ... 1, we aim to study the problem ofuser pro?ling in ego networks. . In this section, we will for-mally introduce ego networks and abstract our problem.First, we formally de?ne users? ego networks. . An egonetwork is a (personal) sub network around an ... friends and those amonghis friends. Here, we assume that an ego network is undi-rected, as many social networks (e.g., Facebook, LinkedIn)are undirected. ... undirected. Our method can be easily extended to han-dle directed ego networks. . As Fig 1 illustrates, a user?s egonetwork can be ... friends, and E? to denote the connectionsamong the friends.In an ego network, , users have attributes (e.g., employer,college, location, ages). We use ...
... missing, and we aim to pro?le each user?s attributes inhis ego networks. ... . Thus, in our setting, we are given a userv0?s ego network G(V,E), where the ego user v0?s attributesare unknown and only ... Figure 1: An Example of an Ego NetworkUser Pro?ling in Ego Network Given a user v0?s egonetwork G(V,E), a list of attributes, ... attribute). However, this assump-tion fails to capture that, in an ego network, , 1) only a fewfriends share the same value with ... assumption used by the exiting methods.We note that, in an ego network, , the ego user?s connec-tions (e.g., e02) have a one-to-one ...
di?erent attributes in dif-ferent ego networks (e.g., student users share their locationswith their classmates, while employee ... circles. Following our insight, we investigatea large collection of real-world ego networks to character-ize 1) the dependency between attributes and circles and ... and connections. Speci?-cally, we conduct a user study to collect ego networks fromLinkedIn, and obtain circles (relationship types) of ego users?friends (connections). ... we canassume that there are K circles in a given ego network, , andthe friends in each circle share the same value ...
... ego user connects(or belongs) to all the circles in his ego network (e.g., col-leagues in Google, college classmates in UIUC), so we ...
... algorithm is polynomialto the number of the nodes in an ego network, , which is verye?cient. Speci?cally, in an iteration, the ?rst ... the best one according to an objective func-tion. Since an ego network is small and our algorithm runse?ciently, it is possible for ...
... we collect his attributes(e.g., employer, college, and location) and his ego network( (including the connections from the ego user to his friendsand ... di?erent colleges, 375 di?erent employers,and 52 di?erent locations, and their ego networks containabout 19K users and 110K connections, among which 8Kare labeled.6.2 ... attributes based on the attributesof the labeled friends and the ego network structure withour algorithm and baseline methods.We use the accuracy to ... on nodes? features ?rst. In oursetting, the graph (i.e., the ego network) ) is given. APi it-eratively estimates every user?s attribute value ...
... a full social network. Here, we adaptit to each user?s ego network. . We can view APc as a base-line, which ?rst ... ?rst applies an existing clustering algorithmto ?nd circles in an ego network and then pro?les the egouser based on discovered circles.Note that, ... be 1) it fails to utilize additional con-nections available in ego networks and 2) the assumptionthat two connected users are likely to ...
... task to identify two attribute-related circles(relationship types) for each given ego network. . They arecolleagues, in which the friends are likely to ... same experiment setting as the previoustask, i.e., in a given ego network, , there are only a few friendswhose attributes are known. ... it with three baseline methods.? RPa pro?les circles in an ego network based on users?attributes. Intuitively, a friend is likely to belong ... on given and pro?led attributes.? RPn pro?les circles in an ego network based on the net-work structure [5]. Intuitively, friends in a ... only on the network structure.? RPan pro?les circles in an ego network based on bothusers? attributes and the network structure [12], which ...
at-tributes based on their ego networks. . We make the followingcontributions to the problem. First, we ... network structure. Forth, we conduct extensive ex-periments based on real-world ego networks to evaluate ouralgorithm. The results show that our co-pro?ling algorithmnot ... J. McAuley and J. Leskovec. Learning to discoversocial circles in ego networks. . In P. Bartlett,F. Pereira, C. Burges, L. Bottou, and ...

Keywords:
ego networks

References:
Everett, M. and Borgatti, S. P. (2005). Ego network betweenness. Social networks (Soc. networks) 27, 1, 31--38.

Full Text:
... for the design of these systems.Keywords ? Empathy; social capital; ego networks; ; online communitiesI. INTRODUCTION Individuals are increasingly establishing social capital ...
... an ego in the quick transmission of information within the ego network [8].We also recorded the number of friends and the number ...
... Media & Society.8. Everett, M. and Borgatti, S. P. (2005). Ego network betweenness. Social networks (Soc. networks) 27, 1 , 31?38.9. Galinsky, ...

Keywords:
ego networks

References:
V. Arnaboldi, M. Conti, A. Passarella, and F. Pezzoni. Analysis of Ego Network Structure in Online Social Networks. In SocialCom, pages 31--40, 2012.
V. Arnaboldi, M. Conti, A. Passarella, and F. Pezzoni. Ego Networks in Twitter: an Experimental Analysis. In Netscicom, 2013.

Full Text:
... 2013, Boston, Massachusetts, USA.Copyright 2013 ACM 978-1-4503-2084-9/13/10 ...$15.00.http://dx.doi.org/10.1145/2512938.2512949.KeywordsOnline social networks, Ego networks, , Personal social rela-tionships1. INTRODUCTIONOnline Social Networks (hereafter OSN) are ...
... a technical point of view we refer thereader to [10].2.1.2 Ego networks and evolutionary psychologyA different approach to the study of social ... at the local properties of personal so-cial networks, often called ego networks. . An ego network( (depicted in Figure 1) is a simple model that describes ... Figure 1) is a simple model that describes the16Figure 1: Ego network modelsocial relationships between an individual, called ego, andall the contacts ... people, called alters - asdefined by the standard notation in ego network analysis.The tie strength between ego and alters is modelled as ... involved in the relationship [4,16].The most important result found on ego networks is thatthere is a limit on the number of alters ... in o?ine and online socialnetworks [4,13,31].Another important result found on ego networks is thatalters around ego form a layered structure with a ... relationships.In anthropology and psychology literature four layers areusually identified in ego networks, , as depicted in Figure 1.The first layer, called support ... cognitive resources. For a complete ethnographic def-inition of the different ego network layers see [25].Ego networks in evolutionary psychology have been stud-ied mainly to understand the ...
... the Dunbar?s number [13]. In [2] and [3] the authorsfound ego network structures in Facebook and Twitter simi-lar to those found in ...
... found that the tie strength is not evenly dis-tributed within ego networks, , but it follows a specific shape,called ?social signature?, characterised ... the network and they donot consider local aspects of the ego networks of the users.Moreover, they consider only the network of social ... study on Twitter [23] revealed that the struc-ture of the ego networks of the users is related, as in o?ineenvironments, to parts ... bythe users into different topics, they found that users withless-constrained ego networks structures (i.e., with accessto structural holes) tend to cover diverse ...
... to carry out a detailed analysis of the dynamicsof Twitter ego networks. .3.3 Detecting humans in the crowdSince our goal is to ...
... leads to sufficiently accurate results [4].Using the standard terms in ego network analysis, we callego a user associated with a profile and ... a social relationship. This definitiongives a ?static? view of the ego networks in the data set,aggregating all the communication of the egos, ... [2?4,13,16]. This allowed us to make aqualitative comparison between the ego network size in ourdata set and that found in other studies ... but is considerably lower thanthe Dunbar?s number found in o?ine ego networks [31]. Thiscould be due to the fact that Twitter is ... in Twitter is still limited [7].To better understand how these ego networks evolve overtime we analysed the time series of the tweets ... the network is active. Instead, foranalysing the dynamics of the ego network structure, wesliced the tweets time series taking snapshots of the ... year each, then assessing the size and the composi-tion of ego networks in each snapshot. We slid the one-yeartemporal window taking steps ... one-yeartemporal window taking steps of one day each, looking athow ego networks change over time.By taking temporal windows of one year we ...
... these different categories of users toshow different behaviours and different ego network proper-ties. Our data set is composed of 63.23% of occasional ... in the figures presented in Section 5, regardingthe composition of ego networks in each one year snapshots(right-hand side plots in Figures 4 ... the axis is equal to the max-imum lifespan of the ego networks in the considered class,minus the duration of the snapshot (one ... consecutive messages within each relationship. Weaveraged the results within the ego networks and then aver-aged for all the ego networks. . While this clearly mixes theproperties of different type of ... theproperties of different type of social relationships for a par-ticular ego network, , it provides a unique index that allowsus to compare ... it provides a unique index that allowsus to compare the ego networks of different classes of users,as explained in detail in Section ... detail in Section 5.After the analysis of the evolution of ego networks ... andpersonal social relationships over time, we measured the sta-bility of ego networks, , assessing the proportion of alters thatusers maintain in their ... are two sets, in our case the one-yearwindows of the ego networks. . The Jaccard coefficient canbe a value between 0 and ... allowed us to determine the ?turnover? thattakes place in the ego networks. . This study is fundamental4Note anyway that we still have ... 0 100 200 300 400Support cliqueDays since first contactFigure 4: Ego networks properties for occasionalusersfor understanding whether people maintain a stable networkof ... define two dis-tinct classes of users: (i) users with structured ego networks, ,showing ego networks with composition and turnover similarto those found in other more ... in other more traditional social networks and(ii) people without structured ego networks, , showing higherturnover.5. RESULTSIn this Section we report the results ...
... only to ?give it a try?, but abandonedit very soon.5.2 Ego networks evolution over time5.2.1 Number of different alters contactedThe first result ... 0 300 600 900 1200Support cliqueDays since first contactFigure 5: Ego network properties for regular userscontacts - daily? depicts the number of ... three classes, re-spectively. This indicates that occasional users have moredynamic ego networks, , with a higher number of new sociallinks added over ... an importantaspect of human social behaviour, indicating high dynamismin the ego networks of the users, that are constantly contact-ing new people rather ... below.To understand how the constant addition of new contactsin the ego networks impacts on the communication level withthe set of existing alters, ... set of alters actively contacted) for thedifferent layers in the ego networks. .5.2.2 Number of alters actively contactedEven though the number of ... time, the number of alters that are activelymaintained in the ego networks does not increase at thesame rate. This fact reveals the ... fact reveals the presence of a turnover strat-egy within the ego networks, , since the new contacts replaceother relationships that are not ... 0 500 1000 1500 2000Support cliqueDays since first contactFigure 6: Ego network properties for aficionadosof the ego network layers are depicted in the right columnof Figures 4 to ... users. Aficionados show a considerable growthin size in all the ego network
... a larger amount of network resources [15],rather than more stable ego networks with stronger andwell-consolidated relationships. Note however that for afi-cionados (i.e., ... at which egos contact new users is nega-tively correlated with ego networks growth rate, indicatingthat users spending a lot of their time ...
... an initial boost of activity,followed by abandonment or gradual decay.5.5 Ego network turnoverFinally, we assessed the stability over time of each layerfor ... calculated the aver-age Jaccard coefficient between separated one-year windowsin each ego network. . To perform this analysis we further re-duced the number ... To perform this analysis we further re-duced the number of ego networks in the data set, since weneeded at least two years ... different layers are reported inTable 2 under the label ?all ego networks? ?. The low valuesof Jaccard coefficient for all the layers ... 0 300 600 900 1200Support cliqueDays since first contactFigure 9: Ego network properties of structured egonetworksauthors found that for the top 20 ... - the turnover islower than for the rest of the ego network. . It is also worthnoting that occasional users show higher ... the higher is the probabilitythat the social relationships in the ego network change dueto turnover.The low values of Jaccard coefficient in the ... Jaccard coefficients considering only users that al-ways maintain a structured ego network, , or, in other words,that show a non empty support ... The results are reported in Table 2 underthe label ?structured ego network? ?. In this case the valueson the Jaccard coefficient for ... the inner layers are stronglyaffected by turnover. The number of ego networks
... other social networks involvingmore traditional and dyadic communications. Remarkably,in structured ego networks the categories of users with longerlifespan have higher values of ... the inner layers. This tells us that users that maintainstructured ego network tend to reinforce their close relation-ships over time, instead of ... results are shown in Figure 9. The activelifespan of these ego networks ranges between 730 and 1, 749days. These are the minimum ... 1, 749days. These are the minimum and maximum active lifetimesof ego networks in our dataset that always presented a non-empty support clique. ... we presented a detailed analysis of the dy-namic processes of ego networks and personal social relation-ships in Twitter. The results indicate that ... aver-age, compared to more traditional social networks, Twit-ter presents smaller ego networks with a high percentage ofweak ties and a really high ... the general behaviour of Twitter users is tomaintain a light-weight ego network formed of weak socialrelationships suitable to maximise the amount of ... of activity and more stable networks. This type ofuser shows ego networks much more similar to those foundin previous analyses of social ... between the properties of their Twitter networks andconventional models of ego networks represents an interest-ing fact, since their behaviour seems to be ...
... V. Arnaboldi, M. Conti, A. Passarella, and F. Pezzoni.Analysis of Ego Network Structure in Online SocialNetworks. In SocialCom, pages 31?40, 2012.[3] V. ...

Title:
Microblogging Personalized Recommendation Based on Ego Networks

Keywords:
microblogging, personalized recommendation, ego networks, homophily, latent factor model

Abstract:
... we propose a novel model of microblogging recommendation based on ego networks, , called EgoRec, which has the ability of capturing personal ... idea of EgoRec is to make recommendation by homophily of ego networks. . We take egos as target users being recommended, while ... ego-alter social relations. We define three main social relations in ego networks: : support relation, affinity relation and focal relation for recommendation. ... detecting latent factor and combining with the homophily factors in ego networks, , it makes EgoRec more personalized. Extensive experiments on real-world ...

Full Text:
... paper, we propose a novel model ofmicroblogging recommendation based on ego networks, , calledEgoRec, which has the ability of capturing personal interestsmore ... key idea of EgoRec is to makerecommendation by homophily of ego networks. . We take egosas target users being recommended, while alters ... and ego-alter social relations.We de?ne three main social relations in ego networks: : supportrelation, af?nity relation and focal relation for recommendation.Moreover, by ... by detecting latent factor and combining with thehomophily factors in ego networks, , it makes EgoRec more per-sonalized. Extensive experiments on real-world ... in ?local? social structures is the goalof the analysis of ego networks [3]. Fig.1(a) shows a socialnetwork graph. It focuses on the ... focuses on the global structure and wholegroup. Fig.1(b) shows an ego network graph, which is a sub-network that focuses on an individual ...
... we propose ego pro?le factor and ego-altersocial relations factor in ego networks as a proxy of homophily.We further de?ne three kinds of ... the microblogging users thatare directly connected to the ego.De?nition 3 (Ego networks) ) Ego networks are socialnetworks consisting of an ego together with alters and ... together with alters and all theties among them. In an ego network, , only ego and users that aredirectly adjacent are included. ...
... we present the details of our EgoRec model.First, we build ego networks based on following relationshipof microblogging, then we propose many factors ... factor and take alistwise ranking algorithm to generate recommendations.A. Building Ego Networks For Microblogging UsersMicroblogging services commonly rely on the basic op-eration ... to the establishment of?following? links. That is, we build users ego networks basedon ?following? relations.Considering an ego network graph, taking the target user(being recommended user) as an ego ... ego are the alters. It is a directed graph.We build ego networks based on ?following? relationship, asshown in Algorithm 1.Algorithm 1 Building ... asshown in Algorithm 1.Algorithm 1 Building ego networksInput: u; userdatabase;relationdatabase.Output: ego networks Eu.1: Createnode (u);2: for i = 1 to u.FolloweeNum do3: ... to de?ne different important socialrelations between ego and alters in ego networks. . In this paper,three kinds of social relations factors were ... a social af?nity measure. For simplicity,we consider the subgraph of ego networks only includingR-friends relationship. Therefore, the subgraph G(V,E) isundirected. Let Gp ...
... interpretedas a form of popularity, and outdegree as gregariousness.On microblogging ego networks, , higher indegree means thatthe ego has much more followers. ... those paths that passthrough v. There are many nodes in ego networks, , and thenodes with high betweenness centrality keep certain stabilityover ...
... demonstrate the effectiveness of the proposed microblog-ging recommendation based on ego networks approaches, we comparethe recommendation results of the following approaches:TIMELINE: This ... and ego-alter social relations. We consider three mainsocial relations in ego networks: : support relation, af?nity relationand focal relation for recommendation. Ego ...
... ego and are takenas the assuming followees. Furtherly, we build ego networks based onthese relationships. It is generated a recommendation list on ...
... paper, we propose the EgoRec model for microbloggingrecommendation based on ego networks. . EgoRec makes recom-mendation by homophily of ego networks. . We quantize homophilybetween individual users by ego pro?le factors ... by detecting latent factor and combiningwith the homophily factors in ego networks, , it makes EgoRec morepersonalized. Our method takes a listwise ...

Title:
Social circle analysis on ego-network based on context frequent pattern mining

Keywords:
ego-network

Abstract:
<p>Effective organization and management of a user's ego- -network is one of the effective means to deal with information ... frequent itemsets mining on users' interactive context, a user in ego- -network can belong to more than one social circles or even ... show that our model can effectively detect social circles on ego- -network.

References:
J. McAuley and J. Leskovec. Discovering social circles in ego networks. arXiv preprint arXiv:1210.8182, 2012.

Full Text:
Social circle analysis on ego- -network based on context frequent pattern miningSocial Circle Analysis on ego- -network based on ContextFrequent Pattern MiningBingying Xu, Lei Deng , Yan ... closed frequentitemsets mining on users' interactive context, a user in ego- -network can belong to more than one social circles or even ... and Subject DescriptorsH.3.4 [Systems and Software]: Information networkGeneral TermsTheoryKeywordssocial circle, ego- -network, , context, frequent pattern mining1. INTRODUCTIONWith the development of IT ... found.In this paper, the discovery and interpretation of socialcircle in ego- -network are researched at the same time. Socialcircle analysis of one ... the same time. Socialcircle analysis of one certain micro-blog user's ego- -network ispresented and explanation the formation of the social circlesare given.2. ...
... frequentitemsets are found to analysis the social circles in one ego- -network. .3. INTERACTION OF MICRO-BLOG3.1 Interactive RelationDivided the behavior of micro-blog ... B, B follows A, thenthey become fans of each other.The ego- -network is such a network,in which the node isuser's friend. And ...
... we proposed a framework tomine all social circles in one's ego- -network , more detail canbe shown in Part 4.4. SOCIAL CIRCLE ... which is obtained in the last offrequent mining process with ego- -network of the user. Theprocess framework of social circle analysis in ...
... users which have common attribute char-acteristics or interactive features in ego- -network, , and theyare also the factors of forming the subset. ... So mining frequentpatterns can not only divide social circles on ego- -network, ,but also can give direct explanation of formation mechanism.But frequent ... itemsets,such as take l greater than or equal to 2.142Given ego- -network G = (V, E), where V is node set and ...
... on the steady state.6. CONCLUSIONSIdentifying the social circles in one's ego- -network and un-derstanding the formation of the social circles can improvemicro-blog ... The paper provides a method to analysis socialcircles in one's ego- -network using context frequent itemsetsmining, combing structure,user pro?les and published con-tent. ...
... 631{640. ACM,2010.[13] J. McAuley and J. Leskovec. Discovering social circlesin ego networks. . arXiv preprint arXiv:1210.8182, 2012.[14] L. W. MO Tong, CHU ...

Title:
Discovering social circles in ego networks

Keywords:
ego networks

Abstract:
... task as a multimembership node clustering problem on a user's ego network, , a network of connections between her friends. We develop ...

References:
J. McAuley and J. Leskovec. 2012. Learning to discover social circles in ego networks. In Neural Information Processing Systems.

Full Text:
... this task as a multimembership node clusteringproblem on a user?s ego network, , a network of connections between her friends. We develop ... Algorithms, ExperimentationAdditional Key Words and Phrases: Community detection, social circles, ego networks, , machine learningACM Reference Format:Julian McAuley and Jure Leskovec. 2014. ... Format:Julian McAuley and Jure Leskovec. 2014. Discovering social circles in ego networks. . ACM Trans. Knowl.Discov. Data 8, 1, Article 4 (February ... date: February 2014.4:2 J. McAuley and J. LeskovecFig. 1. An ego network with labeled circles. The central user, the ego, is friends ...
... problem of circle detection as a clustering problem on her ego network, ,the network of friendships between her friends. In practice, circles ... this task. The firstis the set of edges of the ego network. . We expect that circles are formed by denselyconnected sets ... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:32005], and many circles are hierarchically nested in larger ones ... data.We evaluate our method on a new dataset of 1,143 ego networks from Facebook,Google+, and Twitter, for which we obtain hand-labeled ground ...
... Section 5, we show how to scale the model tolarge ego networks. . We describe the datasets that we construct in Section ... user.The input to our model is a (directed or undirected) ego network G = (V, E), alongwith profiles for each user v ... v ? V . The center node u of the ego network (the ego)is not included in G, but rather G consists ... G consists only of u?s friends (the alters). We definethe ego network in this way precisely because creators of circles do not ... circles do not themselvesappear in their own circles. For each ego network, , our goal is to predict a set of circlesC ...
February 2014.Discovering Social Circles in Ego Networks 4:5school; the corresponding dimension of ?k would then be high ... model of social circles is defined as follows. Given an ego network G and a set ofK circles C = {C1 . ...
... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:7where xi and xj are boolean variables, and Ei, j ... readily handle all but the largest problem sizes typically observedin ego networks: : in the case of Facebook, the average ego network has around 190 nodes[Ugander et al. 2011], whereas the largest ... since the method is unsupervised,inference is performed independently for each ego network. . This means that our methodcould be run on the ... example), as circles are independentlydetected for each user, and the ego networks typically contain only hundreds of nodes.In Section 4, we describe ...
semisupervisedsettings.ALGORITHM 1: Predict complete circles with hyperparameters ?, K.Data: ego- -network G = (V, E), edge features ?(e) : E ? ... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:9In other words, an additional circle will only be added ... setting is realistic if usersconstruct their circles only after their ego networks have already been defined. On theother hand, in settings where ...
... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:11Fig. 2. Histogram of overlap between circles (on Facebook). A ...
... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:13obtain high-quality hand-labeled data, which we have made available.2?4 We ... data, which we have made available.2?4 We were ableto obtain ego networks and ground truth from three major social networking sites:Facebook, Google+, ... Twitter.From Facebook, we obtained profile and network data from 10 ego networks, , consist-ing of 193 circles and 4,039 users. Data was ...
... obtained publicly accessible data. From Google+,we obtained data from 133 ego networks, , consisting of 479 circles and 106,674 users.Data was collected ... least two lists, stopping once we had 1,000 users. The ego networks thatwe obtained range in size from 10 to 4,964 nodes.We ... of the same form.7Taken together, our data contains 1,143 different ego networks; ; 5,541 circles; and192,075 users. The size differences between these ... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:15Fig. 4. Features obtained from Facebook. Each of the 26 ...
... (l ? Ty)). (17)Note that feature descriptors are defined per ego network: : although many thousands ofhigh schools (for example) exist among ... a constant number of dimensionsregardless of the size of the ego network (26 for Facebook, 6 for Google+, 2 for Twitter,as described ... of circles should have commonrelationships to the ego of the ego network. . In this case, we consider the profile tree Tufrom ...
... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:17This measure assigns equal importance to false positives and false ...
... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:19Table I. Methods for Circle and Community DetectionNetwork Node/Edge Overlapping ...
... detail. Figure 6 shows resultsof our unsupervised method on example ego networks from Facebook and Google+.Different colors indicate true/false positives and negatives. ... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:21Fig. 6. Top: Three detected circles on a small ego network from Facebook, compared to three ground-truthcircles (BER ? 0.81). Blue ...
... we examine the problem of adding new users to already-defined ego networks, ,in which complete circles have already been provided. For evaluation, ... For evaluation, we suppress asingle user u from a user?s ego network and learn the model parameters ?? that best fitG {u} ... for 10 random choices of the user u for each ego network in our dataset.As a baseline, we compare the performance of ... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:23Fig. 9. Number of seeds versus accuracy (1 ? BER) ... for each value of K, performance is only evaluatedon those ego networks with at least K ground-truth circles.Figure 9 shows the performance ...
... examines how our algorithm scales with the size of an ego network. . Here weuse the MCMC version of our algorithm described ... LeskovecFig. 10. Running time of our MCMC algorithm for different ego- -network sizes and different values of K (thenumber of circles to ... we fix the feature dimensionality at F = 10 forall ego networks, , using the 10 most common features that appear in ... efficient for small networks(identifying K = 10 circles on an ego network with 1,000 nodes requires around 1 hour),it has the advantage ... that it is easily able to scale to the largest ego networks that areever encountered. For very large networks, the algorithm is ... the best performance of our algorithm is obtained on reasonablysmall ego networks, , although in practice small networks account for the vast ... vast majority ofour data (the data become very sparse for ego networks with more than 300 nodes). Notethat the results for any ... BIC described in Section 3.1.Performance degrades only slightly for large ego networks, , although it remains anopen question whether this is due ... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:25Fig. 11. Accuracy of our MCMC algorithm in terms of ... example, if a user has a Stanford circle in his ego network, , it is highly likelythat users belonging to that circle ... a user?s Stanford circle might simply bethe intersection of his ego network with that community. Studying such models is anappealing avenue for ...
... ameliorated withmore data.We developed algorithms that scale to the largest ego networks that we encountered;however, we find that the best performance occurs ... we encountered;however, we find that the best performance occurs on ego networks with up to a fewhundred nodes but degrades significantly for ... manually. We havedesigned an algorithm to automatically detect circles in ego networks, , which we eval-uated on a dataset of 1,143 ego networks and 5,541 ground-truth circles obtained fromFacebook, Google+, and Twitter. In ... 1, Article 4, Publication date: February 2014.Discovering Social Circles in Ego Networks 4:27J. Chang and D. Blei. 2009. Relational topic models for ...
... and J. Leskovec. 2012. Learning to discover social circles in ego networks. . In Neural InformationProcessing Systems.M. McPherson. 1983. An ecology of ...

Keywords:
Ego Networks

Abstract:
... interested to explore and understand their personal networks (called the ego networks) ) and how their contacts are associated with different social ... we propose an enhanced multi-layered Chord layout to show people' ego networks and the associated social circles in interactive visual forms. This ...

References:
Ezaiza, H., Humayoun, S. R., AlTarawneh, R., and Ebert, A. 2016. PerSoN-Vis: Visualizing Personal Social Networks (Ego Networks). CHI Extended Abstract '16, ACM.
Freeman, L. C. 1982. Centered Graphs and the Structure of Ego Networks. Mathematical Social Sciences, v. 3, 291--304.
Mcauley, J. and Leskovec, J. 2014. Learning to Discover Social Circles in Ego Networks. TKDD, 8(1), Article 4.

Full Text:
... interested to explore and understand their personal networks (called the ego networks) ) and how their contacts are associated with different social ... we propose an enhanced multi-layered Chord layout to show people? ego networks and the associated social circles in interactive visual forms. This ... set of provided interaction and filtering options. Keywords Social Networks; Ego Networks; ; Personal Visualization; Circle Discovery; Chord Layout. CCS Concepts ? ... a personal network inside a social network is called the ego network [3] or the personal social network [8] of the underlying ... researchers have talked about the importance of studying and analyzing ego networks [3, 5, 8]. In general, visual representations enable human beings ... to having a strong visual system [7]. Therefore, visualizing people? ego networks in intuitive and interactive forms would help them to explore ... visual social tool called PerSoN-Vis [4]. It visualizes intuitively people? ego- -networks through the node-link layout and the Chord layout. PerSoN-Vis enables ... with multiple social circles. Visual exploration of social circles in ego networks requires to visualize two forms of relations, i.e., the relations ... visualization for a better and interactive exploration of the underlying ego network and the associated social circles. 2. THE MULTI-LAYERED CHORD LAYOUT ... This data is related to 10 different users, i.e., 10 ego networks. . The nodes in these ego networks, , representing friends of these 10 users on Facebook, associated ...
... these collected features. In Figure 1, we show a user?s ego network in two visual representations, i.e., the standard Chord layout (Fig. ... 1a) and our proposed multi-layered Chord layout (Fig. 1b). This ego network is consisted of 59 nodes (i.e., the contacts), 146 edges ... June 07-10, 2016, Bari, Italy ACM 978-1-4503-4131-8/16/06. DOI: http://dx.doi.org/10.1145/2909132.2926072 314underlying ego network and their relations. The color of the relation indicates the ... relation indicates the source (green) to destination (red). Figure 1. Ego network of a user: (a) using the standard Chord layout (b) ... circle represents the main social circles, associated to the underlying ego network. . These social circles (e.g., country, education, music, social cause, ... filtering and interaction options for better exploration of the underlying ego network and the associated social circles. For example, user can select ... select a particular node representing a contact in the current ego network. . This results in making all the non-associated nodes and ... nodes; however, these non-associated nodes remain grey. Additionally, the underlying ego network can be filtered based on selecting more than one social ... multi-layered Chord layout targeting exploration of social circles in people? ego networks. . This would help them to better understand and to ...
... R., and Ebert, A. 2016. PerSoN-Vis: Visualizing Personal Social Networks (Ego Networks) ). CHI Extended Abstract '16, ACM. [5] Freeman, L. C. ... Freeman, L. C. 1982. Centered Graphs and the Structure of Ego Networks. . Mathematical Social Sciences, v. 3, 291-304. [6] Freeman, L. ... and Leskovec, J. 2014. Learning to Discover Social Circles in Ego Networks.

Keywords:
ego networks

References:
Everett, M. and Borgatti, S. P. (2005). Ego network betweenness. Social networks (Soc. networks) 27, 1 (2005), 31--38.

Full Text:
... interfaces and presentation (e.g., HCI): Miscellaneous.General TermsHuman Factors.KeywordsSocial networks analysis, ego networks, , social capital, personality traits.1. INTRODUCTIONIn our work we seek ...
... of friends and number of ties between them in the ego network. . The fewer the ties between the ego?s friends, the ... the ego?s friends, the greater the effective size of the ego network (for the exact definition the reader may refer to [2]).? ... an ego in the quick transmission of information within the ego network [5].In addition we examine the following metrics in relation to ... social capital:? Degree centrality: The number of friends in the ego network. .? Isolated friends: The number of friends in the ego network with no other common friend with the ego.? Transitivity : ... probability that any two friends of an individual in the ego network are in turn friends.? Density : The ratio of the ... links and the total number of possible links in an ego network. .2.2.2 Social CapitalBridging and bonding social capital was measured with ...
... of structural holes (measured by betweenness and constraint) in individuals? ego networks and bridging social capital. This is in agreement with the ...
... Communication, 12(4), 143-1168.5. Everett, M. and Borgatti, S. P. (2005). Ego network betweenness. Social networks (Soc. networks) 27, 1 (2005), 31?38.6. Granovetter, ...

Keywords:
ego network

References:
Everett. M. and Borgatti. S.P.; Ego network betweenness, Social Networks 27, 31--38, (2005).

Full Text:
... nodes in the network in aspect of information propagation.KEYWORDSInformation propagation, ego network, , betweenness centrality, close-ness centrality, degree distribution, random walkACM Reference ... are found out based on the four important graphproperties i.e., ego network, , betweenness centrality, closeness cen-trality and degree distribution.3 ANALYSIS OF ... distribution.3 ANALYSIS OF NETWORK3.1 Ego NetworkEgo speci es the nodewithmaximumdegree and ego network con-verts the network converging around the ego node[2]. Degree ofall ... a er that vertex namewithmaximum degree is obtained and feed into ego network. . e ver-tex contains highest degree in a network considered ...
... eoretical Computer Science, 412, 2602-2610, (2011).[2] Evere . M.and Borga i. S.P.; Ego network betweenness, Social Networks 27 , 31-38,(2005).[3] Brandes. U.; A faster ...

Title:
Evolution of Ego-networks in Social Media with Link Recommendations

Keywords:
ego-networks

Abstract:
<p>Ego- -networks are fundamental structures in social graphs, yet the process of ... this matter, we analyze the complete temporal evolution of 170M ego- -networks extracted from Flickr and Tumblr, comparing links that are created ... have been algorithmically recommended. We find that the evolution of ego- -networks is bursty, community-driven, and characterized by subsequent phases of explosive ...

References:
V. Arnaboldi, M. Conti, M. La Gala, A. Passarella, and F. Pezzoni. Ego network structure in online social networks and its impact on information diffusion. Comp. Comm., 2016.
V. Arnaboldi, M. Conti, A. Passarella, and F. Pezzoni. Analysis of ego network structure in online social networks. In SocialCom, 2012.
V. Arnaboldi, M. Conti, A. Passarella, and F. Pezzoni. Ego networks in twitter: an experimental analysis. In InfoCom, 2013.
V. Arnaboldi, R. Dunbar, A. Passarella, and M. Conti. Analysis of co-authorship ego networks. In Network Science. Springer, 2016.
A. Biswas and B. Biswas. Investigating community structure in perspective of ego network. Expert Syst. Appl., 2015.
A. De Salve, M. Dondio, B. Guidi, and L. Ricci. The impact of user's availability on on-line ego networks: a facebook analysis. Comp. Comm., 2016.
L. C. Freeman. Centered graphs and the structure of ego networks. Math. Soc. Sci., 1982.
S. Gupta, X. Yan, and K. Lerman. Structural properties of ego networks. In SBP-BRiMS. Springer, 2015.
J. Mcauley and J. Leskovec. Discovering social circles in ego networks. TKDD, 2014.
S. A. Muhammad and K. Van Laerhoven. Duke: A solution for discovering neighborhood patterns in ego networks. In ICWSM, 2015.

Full Text:
... on this matter,we analyze the complete temporal evolution of 170M ego- -networks extracted from Flickr and Tumblr, comparing linksthat are created spontaneously ... be used to improve the effectiveness of socialrecommender systems.KeywordsNetwork evolution, ego- -networks, , link recommendation, groups,communities, social media, Tumblr, Flickr1. INTRODUCTIONEgo-centric social ... link recommendation, groups,communities, social media, Tumblr, Flickr1. INTRODUCTIONEgo-centric social networks (ego- -networks) ) map the in-teractions that occur between the social contacts ... has been conducted to unveil thestructure and evolutionary dynamics of ego- -networks [4, 30].In an online context, people expand their social circles ... We do so by analyzingthe full longitudinal traces of 170M ego- -networks from Flickrand Tumblr ( 3), answering several open research questionsabout the ... of our key findings are:? The backbone of a typical ego- -network is shaped withinthe initial month of a node?s activity and ... exploration strategy.? The link recommender system skews the process of ego- -network construction towards more popular contacts butat the same time restraining ...
... becrucial features to improve link prediction performance, asthe criteria of ego- -network expansion vary as the ego growsolder. In a second experiment, ... has been measuredin several online social graphs including Flickr [33, 57].Ego- -networks. . To date, not much research has been con-ducted on ... neighborhoodsin time. Research done by Aranboldi et al. has lookedinto ego- -networks of online-mediated relationships includingthe Facebook friendship network [4] and the ... of comparing the properties ofthe global network with those of ego- -networks, , recent stud-ies found that local structural attributes are characterized ... have been proposed to dis-cover social or topical sub-groups within ego- -networks [53,37, 40, 13], but with little attention to the dynamics ... the few studies touch-ing upon the the temporal evolution of ego- -networks, , usinga dataset of Skype contacts [30]. They find that ...
... of linkrecommendation systems, a key open question is how theyaffect ego- -network diversity.3. DATASET AND PRELIMINARIESWe study two social media platforms that ... the whole lifespan of the platforms.3.3 Concepts and notationGraph and ego- -network. . Consider a follower graph Gcomposed by a set of ... node j if i has followed j at any time.The ego- -network Gi of node i ? N is the subgraph inducedby ... E| j ? ?out(i) ? l ??out(i)}. Note that the ego- -network does not include thelinks between the ego i and its ...
... ?t(i) and its degree as kt(i). When studyingthe evolution of ego- -networks in isolation, we will considertime on a discrete scale where ... each event corresponds to thenth node being added to the ego- -network. . We will use theletter n to denote time passing ... 600 800 1000n0.00.51.01.52.02.53.0network distanceTumblrTumblr> 1000FlickrFigure 4: Left: average number of ego- -network links vs. numberof nodes; best fitting power-law exponent is reported ... is added;the black dotted line is obtained considering only Tumblr ego- -networks with at least 1000 nodes.at 5000 because the platform imposes ... component [34]. It is unknown whether thesame properties hold at ego- -network level.Q1: How do density, diameter, and component struc-ture evolve as ... How do density, diameter, and component struc-ture evolve as the ego- -network grows?Like global networks, ego- -networks become denser in time.Ego-networks obey a densification power law, for ...
... addition of recommended nodes has a different ef-fect on the ego- -network expansion, compared to sponta-neous ones. New recommended contacts tend to ... sponta-neous ones. New recommended contacts tend to be closerto existing ego- -network members. At fixed network size, theaddition of new recommended nodes ... networkdistance 5% less, on average, than a spontaneous node addi-tion. Ego- -networks ... with at least a recommended node havesmaller network distance than ego- -networks that grew fullyspontaneously; this difference varies with the size, being ... to spawn a new component disconnected from the restof the ego- -network (Figure 5, right).Accounting for amalgamation effects. To explore evo-lutionary trends ... 5, right).Accounting for amalgamation effects. To explore evo-lutionary trends of ego- -networks, , we rely on aggregate anal-ysis: an indicator is measured ... rely on aggregate anal-ysis: an indicator is measured on an ego- -network when itsnth node is added and then it is averaged ... node is added and then it is averaged across all ego- -networks. . Trends are discovered as n grows (e.g., diameterin Figure ... usingthe full dataset and the latter considering only the subsetof ego- -networks that reached at least size nmax. The resultsare only slightly ... ego (common neighbors) at the time of their inclusionin the ego- -network. .100 101 102 103n10-1100101102103common neighborsTumblrTumblr RecsFlickr100 101 102 103n103104105106107108109pref. attachment101 ... Average similarity and popularity indicators of nthnode added to the ego- -network: : common neghbors between theego and the newly added nodes, ... (kout(ego) kin(newnode)), and indegree of the new node.ameter evolution for Tumblr ego- -networks that reached atleast size 1000 is very similar to the ... popularity and similarity impact the selec-tion of new nodes in ego- -networks, , and how their relativeimportance varies in time.Q2: How do ... time.Q2: How do the criteria of neighbor selection changeas the ego- -network grows?We select two simple (yet widely-used) proxies of popu-larity and ...
... to the ego?s neighbors.As new nodes are added to the ego- -network, , the number oftheir common connections with the ego naturally ... the in-terval [100, 1000] and decreasing otherwise. The popularityof new ego- -network members, computed as their indegree inthe social network, decreases as ... computed as their indegree inthe social network, decreases as the ego- -network grows. Asummary of all the indicators is given in Figure ... aim to learnhow often and in which phases of the ego- -network life usersselect new neighbors.Q3: When do ego- -networks expand?To measure how much node additions to an ego-networkare concentrated ... a sessionstarts when a new node is added to the ego- -network and1150 20 40 60 80 100ego age (days)0.000.050.100.150.200.25ratio of nodes ... first 100 daysof the ego?s life, relative to the final ego- -network size. Only nodeswho have created links for at least 6 ... shown). This suggest that nodes tend tobuild most of their ego- -network in the first stages of theirlife. To confirm that, we ...
... size TumblrFlickrFigure 11: Left: distribution of number of communities in ego- -networks. ... . Right: average size of communities as they appear inthe ego- -network; ; 95% confidence intervals are shown.poral granularity found that the ... the role ofthese communities in the graph evolution.Q4: Is the ego- -network growth driven by the bound-aries of its communities?The ego- -network may follow a depth-first expansion pat-tern with respect to communties, ... family members and then thoseof a photography club. Alternatively, the ego- -network mayexpand either breadth-first, picking new nodes in a round-robin fashion, ... assume a static, hard partitioning ofnodes in communities. For all ego- -networks, , we computenon-overlapping communities6 at time t = Tend (the ... data), heuristically filtering out smallego-networks with less than 5 members. Ego- -networks areoften composed by few communities, rarely more than 10(Figure 11, ... 40 60 80 100n0.00.20.40.60.81.0inversion scoreFlickrRealShuffledFigure 12: Average inversion score of ego- -network communitiesas new nodes are added, compared to a randomized null-model.0 ... Average likelihood (over random chance) that thenth node in a ego- -network belongs to the kth community.sequence of community ranks R (e.g., ... a community half of whose members has beenadded to the ego- -network by time t comes temporally beforeany other whose majority of ... beforeany other whose majority of nodes are still outside the ego- -network at the same time t. If exploration of communitiesis purely ... we plot the average in-version score of R against the ego- -network size. To accountfor the community size heterogeneity, we compare it ... measuring the prob-ability that the nth node added to the ego- -network belongsto the kth community in the ranking, normalized by theprobability ...
... same and havebeen added in the same order to their ego- -networks. . Eachmatching group is then split in two subgroups: a ...
... link recommendations that adaptto the current evolutionary stage of the ego- -network couldkin PA CN Jaccard0.0000.0050.0100.0150.020normalized feature value spontaneous linkskin PA CN ...
... content; thatinformation would help to further detail the dynamics of ego- -network expansion with respect to other dimensions such astopical similarity between ... Arnaboldi, M. Conti, M. La Gala, A. Passarella, andF. Pezzoni. Ego network structure in online social networksand its impact on information diffusion. ... V. Arnaboldi, M. Conti, A. Passarella, and F. Pezzoni.Analysis of ego network structure in online social networks.In SocialCom, 2012.[5] V. Arnaboldi, M. ... Arnaboldi, R. Dunbar, A. Passarella, and M. Conti.Analysis of co-authorship ego networks. . In NetworkScience. Springer, 2016.[7] L. Backstrom, D. Huttenlocher, J. ...
... A. Biswas and B. Biswas. Investigating communitystructure in perspective of ego network. . Expert Syst. Appl.,2015.[14] V. D. Blondel, J.-L. Guillaume, R. ... Guidi, and L. Ricci. Theimpact of user?s availability on on-line ego networks: : afacebook analysis. Comp. Comm., 2016.[20] R. Dunbar. Grooming, gossip, ... K. Van Laerhoven. Duke: A solutionfor discovering neighborhood patterns in ego networks. . InICWSM, 2015.[41] G. Palla, A.-L. Baraba si, and T. Vicsek. ...

Keywords:
ego networks

Abstract:
... of the centrality metrics in populated networks the concept of ego networks is exploited where nodes are not required to exchange information ...

References:
EVERETT, M., AND BORGATTI, S. P. Ego network betweenness. Social networks (Soc. networks) 27, 1 (2005), 31--38.

Full Text:
... complexity of the centrality metrics inpopulated networks the concept of ego networks is exploited wherenodes are not required to exchange information about ...
... of this paper is a new for-warding metric based on ego network analysis to locally determinea node?s centrality within the network and ...
... For this reason the concept of?ego networks? has been introduced. Ego networks can be definedas a network consisting of a single actor ... to (alters) and all the links among those al-ters. Consequently, ego network analysis can be performed locallyby individual nodes without complete knowledge ... network [25]. Degree centrality can easily be mea-sured for an ego network where it is a simple count of the numberof contacts. ... is uninformative in an ego net-work, since by definition an ego network only considers nodes towhich the ego node is directly related ...
... solely onlocal calculations.5.1 Betweenness CalculationBetweenness centrality is calculated using an ego network repre-sentation of the nodes with which the ego node has ...

Title:
Visualization of evolving large scale ego-networks

Keywords:
ego-networks

Abstract:
... continuous stream and represent it by streaming the egocentric networks (Ego- -Networks) ) for particular nodes. We propose a non-standard node forgetting ... aim of these techniques is the visualization of large scale Ego- -Networks

References:
DeJordy, R., and Halgin, D. Introduction into ego network analysis, 2009.

Full Text:
... continuous stream and represent it by stream-ing the egocentric networks (Ego- -Networks) ) for particularnodes. We propose a non-standard node forgetting factor ... executed with an ordinarycommodity machine.2. RELATEDWORKSeveral studies have already addressed Ego- -Networks. . Thisfield covers and relates with many varied subjects, from ... Social Networks ismade and a full chapter is dedicated to Ego- -Networks. . Han-neman et al. define ?Ego? as an individual ?focal? ... holes? create en-trepreneurial opportunities for information access, timing,referrals and control. Ego- -Networks analysis provides ananswer to this sensible information or properties that ... section we describe the algorithm used for stream-ing. For the Ego- -Networks streaming representation of thedata we used the landmark window implementation ... alternative to full networks SNAmodel. Fig. 1 shows the example Ego- -Network
... in thelandmark window. The algorithm returns the visualizationof the evolving Ego- -Network over time. New connectionsarising from the central node or their ... node or their 1st order connectionsare plotted in the screen.3.1 Ego- -Networks with forgetting factorEgo-Networks with node forgetting factor algorithm, us-ing streaming ... if the method results were represen-tative of the original data Ego- -Network as the data stream-ing and forgetting factor value evolved over ... 10 at mid-night and reaches its peak atmid-day with 280.4.2 Ego- -Networks with forgetting factorFig. 3 illustrates the test of parameters with ... 3 illustrates the test of parameters with the sameanonymous number?s Ego- -Network presented in Fig. 1. Theupdate time period was set to ... of 0.6.4.2.1 Node and Edges counter variation with timeFigure 3: Ego- -Network with forgetting factor961After the previous experiment we randomly selected threephone/nodes ...
... for nodesand edges. The variation regarded time and comparison be-tween Ego- -Network with and without forgetting factor forthe same update periods. It ... and by changing only the update time period parameter.Figure 4: Ego- -Network nodes counterFigure 5: Ego- -Network edges counterIt is known that the values of both counters ... the values of both counters increase withtime for the original Ego- -Network without forgetting factor.In Fig.4 and Fig. 5, the horizontal lines ... centred networks and ulti-mately, by using forgetting factor in the Ego- -Network out-put, we propose a novel streaming method for this type ... con-sider and prove this to be an effective visualization methodfor Ego- -Networks SNA.Finally, we conclude that this method for evolving net-works visualization ... visualization and analysis is a light method to visu-alize massive Ego- -Networks. . Thus, the simulation of a datastream and the visualization ... leveraging telecommunication systems andenabling the visualization of real time evolving Ego- -Networks. .6. ACKNOWLEDGMENTSThis work was supported by Sibila and Smartgrids re-search ...

Keywords:
ego networks

References:
Linton C. Freeman. 1982. Centered graphs and the structure of ego networks. Math. Soc. Sci. 3, 3 (1982), 291--304.

Full Text:
... Words and Phrases: Gender and age, demographic prediction, node attributes, ego networks, ,social tie and triad, mobile communication, mobile phone data, computational ...
... and demographic characteristics in the perspectiveof micro-level network structures, including ego networks, , social ties, and social triads.We also use a null ... the results in Section 4.3.1. Social Strategies on Ego NetworksAn ego network of one person is defined by viewing himself or herself ... Figure 1 presents an illustrativeexample of the evolution of one?s ego network. . We first examine the characteristics ofthe central node (ego) ... node (ego) and then the distributions of this ego?s friends (ego network) )with respect to their demographic profiles.Ego.We present a basic correlation ... are observed at a 95% confidence interval.Ego Networks. With the ego network of each user, we study the demographic ho-mophily on both ... 2017.35:8 Y. Dong et al.Fig. 5. Friends? demographic distribution in ego networks. . X-axis: (a) the age of a female ego in ...
... the shuffled results and z-score plots. Specifically, the results on ego networks areshown in Figures 9 and 10. The shuffled results and ...
... namely nonstruc-tural attribute features and structural features. Specifically, given an ego network withone central user v and her/his direct friends, we extract ... to model the demographic distribution of v?s direct friendsin her/his ego network, , including the number of connections to female, male, young,young-adult, ... number of senior contacts#v-FF-triangles number of FF-v triangles in v?s ego network# #v-FM-triangles number of FM-v triangles in v?s ego network# #v-MM-triangles number of MM-v triangles in v?s ego network# #v-AA-triangles number of AA-v triangles in v?s ego networkCircle attributes ... ego networkD: senior #v-BC-triangles number of BC-v triangles in v?s ego network# #v-BD-triangles number of BD-v triangles in v?s ego network# #v-CC-triangles number of CC-v triangles in v?s ego network#
triangles in v?s ego network# #v-DD-triangles number of DD-v triangles in v?s ego networkTable III ...
... 066103.Linton C. Freeman. 1982. Centered graphs and the structure of ego networks. . Math. Soc. Sci. 3, 3 (1982),291?304.Huiji Gao, Jiliang Tang, ...